Available online at www.sciencedirect.com
Journal of Science and Medicine in Sport 12 (2009) 352–356
Original paper
Arm–leg coordination in recreational and competitive
breaststroke swimmers
Hugues Leblanc, Ludovic Seifert ∗ , Didier Chollet
University of Rouen, France
Received 17 April 2007; received in revised form 29 December 2007; accepted 8 January 2008
Abstract
The aims of this study were to assess the durations of the different arm and leg stroke phases (propulsion, glide, and recovery) and
the temporal arm–leg gaps between 12 competitive and 12 recreational breaststroke swimmers. The mean ages and best times for a 50-m
breaststroke were, respectively, (recreational: 16.9 ± 1.6 y; 49.55 ± 3.38 s; competitive: 16.2 ± 1.5 y; 33.85 ± 1.96 s). Each swimmer was
required to swim 2 × 25-m breaststroke at two different paces (slow and sprint) while being videotaped by two underwater cameras (frontal
and lateral views). At the same given speed, recreational swimmers used no glide phase which increased the relative contribution of their
recovery and propulsive phases. This was mainly caused by the superposition of their leg extension and the second part of their arm recovery,
indicating a technique with no glide time between the arm recovery and the leg extension. In terms of phase duration, the recreational swimmers
spent more time in arm recovery and in propulsive phases. Furthermore, it was observed that for a comparable increase of swimming speed
(recreational: 23.3%, competitive: 22.6%), competitors switched from a glide to an overlapped coordination while recreational swimmers
adopted an overlapped technique whatever the swimming speed. As a result, the relative time spent in propulsive phases did not change in the
recreational group, but increased by 27.2% in the competitive one. In a swimming developmental program, particular emphasis should be put
on arm–leg coordination drills, when considering the breaststroke.
© 2008 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
Keywords: Swimming; Motor skills; Movement; Coordination
1. Introduction
Competitive breaststroke is characterized by the underwater recovery of the limbs. To face this constraint, the
swimmers simultaneously recover their arms and legs, and
then the sequence leg kick, glide, arm pull and in-sweep follows. According to the glide time, three kinds of coordination
are observed1 : (1) glide where the body stays fully extended
and streamlined before the arm catch, (2) continuous where
the arm catch takes over just as the leg kick is completed,
(3) overlapped where the arms start their catch and outward
motion before the completion of the leg kick. The individual characteristics and the swimming speed are two factors
that can influence the arm–leg coordination of a swimmer.
For example, Sanders2 found that the glide phase could vary
from 0 to 0.55 s per cycle at race pace in male elite swim∗
Corresponding author.
E-mail address: ludovic.seifert@univ-rouen.fr (L. Seifert).
mers. Concerning the swimming speed, glide coordination
has been observed in 200-m events, whereas during shorter
races (100 and 50 m), continuous or overlapped coordination
is more likely used.3–5 This adaptation eventually allows the
swimmer to swim at a higher stroke rate and speed.
The coordination patterns of competitive breaststrokers
can give knowledgeable information to the swimming coach,
but they give [because the subject is plural: patterns] no
hints to understand the coordination of beginners. Up-to-date
data are scarce on this topic. In a former electromyographic
study,6 the arm pulling phase of unskilled young breaststroke
swimmers was associated with noticeable discharges of the
rectus femoris muscle, which inserts on the ilium and on the
patella and is a hip flexor. It was concluded that the arm pull
occurred at the same time as the leg recovery. In other words,
the swimmers were simultaneously doing two contradictory
actions—from one hand, the leg recovery, which caused a
drop of velocity, and from the other hand, the arm propulsion. The authors, however, did not make an analysis of the
1440-2440/$ – see front matter © 2008 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsams.2008.01.001
H. Leblanc et al. / Journal of Science and Medicine in Sport 12 (2009) 352–356
stroke phases or investigate at different swimming speeds.
Therefore, this study was twofold: (i) to make an accurate
assessment of the temporal aspects of the arm–leg coordination of recreational swimmers; (ii) to analyze the evolution
of the arm–leg coordination under different speed conditions.
Within this frame, two hypotheses can be made: firstly recreational swimmers are expected to perform their leg kick while
recovering their arms and to do their arm pull while recovering their legs. Secondly, their coordination pattern is not
expected to evolve from a glide to a continuous or overlapped mode like skilled swimmers do, essentially because
they feature no glide phases during their stroke cycle.
2. Methods
Twenty-four male swimmers participated in this study
after giving their written consent. Twelve competitors
swam at a French regional level, and twelve recreational
swimmers were high school students with no particular
background in swimming. The main characteristics of the
subjects were (for competitors and recreational swimmers,
respectively) age: 16.2 ± 1.5 y, mass: 71.2 ± 6.9 kg, height:
176.5 ± 7.8 cm, best time on 50 m: 33.85 ± 1.96 s; which
represented 77.3 ± 4.4% of the short course male world
record on January 1st/2007; and 16.9 ± 1.6 y., 69.4 ± 10.8 kg,
174.6 ± 11.5 cm, 49.55 ± 3.38 s, 50.9 ± 8.2% (percentage of
world record = (swimmer time/world record) × 100). No significant differences were found between the two groups
for age, mass, and height. Recreational swimmers had
significantly lower best times on the 50-m breaststroke
(F1,22 = 193.40), which represented a smaller percentage of
the world record (F1,22 = 226.56). The F values given here
represent the statistical distribution of the ratio of two variances. It was significant if F > 4.30.
Before the trials, each swimmer was randomly assigned a
passing order. Then, each of them was asked to swim at an
imposed velocity over a set of 2 m × 25 m (slow then sprint
paces), with 5 min rest in-between. For competitors, the slow
pace corresponded to a 400-m breaststroke (their coaches
provided the targeted time in reference to training practice).
Knowing the performance level of the two groups, the 400m speed of the competitors was expected to match the sprint
speed of the recreational swimmers. After each trial, all swimmers were informed of their performances. Competitors were
asked to swim within ±2.5% of their targeted time. Beginners were only asked to swim their slow trial about 20% over
their sprint time as recorded during a 25-m pre-test. The trials were monitored by two experienced timers who controlled
the stroke rate and the speed with a stopwatch and a Seiko
Base 3-frequency-meter to validate each trial. If this was not
the case, the subject had to repeat the trial.
Two Samsung SC107 digital camcorders (DV format)
were connected via an AV/DV analogical input to two underwater cameras (respectively, from a frontal and an 11-m side
view). A third camcorder (Canon Obtura DV) placed on the
353
pool deck videotaped and timed the swimmers over a distance
of 12.5 m (between the 10 m and the 22.5 m marks made on
the pool edges), enabling calculation of the average swimming speed and stroke rate. The stroke length was computed
from this speed and the stroke rate values. A flash light was
used to synchronize the pictures, with an accuracy of 1/66 s.
After being downloaded on a PC, the pictures were analyzed with Dartfish Prosuite 4.0® software (Atlanta, GA).
Underwater views could be mixed and synchronized to enable
the data treatment. The acquisition rate was of 66 frames s−1 .
These phases have been previously described.5 Briefly,
for each pair of limbs, three phases were defined: propulsion,
glide, and recovery. The leg in-sweep was also characterized
as the time elapsed between the end of the leg extension and
their joined position.
To obtain consistent measurements, the graphic tools of
the software were used. For the arm phases, a vertical line
which passed through the shoulder profile axis was drawn to
have a fixed reference on the swimmer’s body. The leg/thigh
angle was measured to determine the leg maximal extension
and end of recovery. The feet distance was checked on the
frontal view to determine the leg in-sweep phase.
Four temporal gaps between the arm and leg actions were
defined:
T1: From the end of the leg in-sweep to the beginning of the
arm propulsion.
T2: From the beginning of leg recovery to the beginning of
the arm recovery.
T3: From the end of leg recovery to the end of the arm
recovery.
T4: From the leg extension to the end of the arm recovery.
T1 was calculated in order to quantify the time during which
the body is fully extended and glides. T2 and T3 aimed
to quantify the arm–leg coordination during the recovery phase. T4 was calculated because it evaluated the
synchronization of the extension of the two pairs of
limbs.
The propulsive index was computed by doing the sum of
the leg and the arm propulsion respective durations.5 If a
propulsive phase of one pair of limbs overlapped the recovery
phase of the other pair, then the overlapped duration was
subtracted from the propulsive phases.
3. Data analysis
Standard statistical procedures were applied to compute
means and standard deviations. The normality of the distributions (Shapiro–Wilk test) and the homogeneity of the
variances were controlled, allowing the use of parametric
statistics. Two-way ANOVA on repeated measures were computed (among subject factor: skill level, within subject factor:
swimming speed). Post hoc comparisons were made by using
the Bonferroni test. The level of significance was set at
p < 0.05.
354
H. Leblanc et al. / Journal of Science and Medicine in Sport 12 (2009) 352–356
As no significant difference was found between the sprint
speed of recreational swimmers, and the slow speed of competitive swimmers, it was chosen to compare the two groups
at the same given speed, while facing a comparable hydrodynamical constraint.7
4. Results
At the same given swimming speed, competitors covered
a significantly longer distance per cycle while maintaining a
lower stroke rate (SR) (p < 0.05) (supplementary Table 1).
The arm propulsive (relative values) and recovery (relative
and absolute values) phases were significantly longer in the
recreational group (p < 0.05). By contrast, the arm glide phase
appears significantly shorter in this group, both in absolute
and relative values (p < 0.05).
No significant difference was found regarding the leg
propulsive phase. As expressed in relative values, the recreational group spent a significantly longer time performing its
leg recovery. The leg glide was significantly smaller in the
recreational group (relative values) (p < 0.05). The duration
of the leg in-sweep was not significantly different between
the two skill levels (supplementary Table 2).
In the recreational group, T1 which expressed the
glide time, appeared to be significantly shorter (p < 0.05)
in the recreational group (relative and absolute values)
(supplementary Table 3). No difference was noticed for T2,
measuring the time interval between the beginning of the
arm and leg respective recoveries. The time interval measured at the end of the arm and leg recovery (T3) showed
a significantly greater negative (p < 0.05) value in the recreational group. T4 measuring the time interval between the arm
and leg respective extensions was smaller in the recreational
group (relative and absolute values). Finally, the propulsion
index of the recreational swimmers was significantly smaller
in absolute values.
Recreational and competitive swimmers performed the
sprint trials at a significantly (p < 0.05) greater speed, with
a higher SR and a shorter stroke length (SL) (supplementary
Table 1). The percentage of increase of speed between the
slow and the sprint trials were not significantly different
between the two groups (respectively, 23.34% ± 0.09 for
recreational swimmers and 22.56% ± 0.06 for competitive
swimmers).
The arm propulsion phases significantly decreased in
recreational as in competitors (relative values, p < 0.05).
However, in the recreational group, this was also verified
in absolute values. The gliding phase of the arm significantly
decrease in the competitive group (p < 0.05), but remained
unchanged in the recreational group, whatever the pace. The
arm recovery phase significantly decreased in real terms in the
recreational swimmers. The leg propulsion phase showed a
similar evolution than that of the arm. Competitive and recreational swimmers significantly diminished their leg glide
with the increase of speed (p < 0.05). For both groups, the leg
recovery phase significantly decreased but only in absolute
value (p < 0.05) (supplementary Table 2).
In both groups, the gliding parameter T1 duration was significantly smaller at sprint speed. T2, T3 and T4 parameters
did not evolve with the speed increase in any of the groups
(p < 0.05). Only T4 as expressed in percentage significantly
increased in the competitive group.
In the recreational group, the propulsive index remained
constant in relative value, but statistically significantly
decreased in absolute terms (p < 0.05). Inverse conclusions
are noted in the competitive group (supplementary Table 3).
5. Discussion
5.1. Skill level comparison at a given speed (sprint pace
for recreational swimmers vs. slow pace for competitors)
The first research hypothesis was partially confirmed:
recreational swimmers perform their arm recovery while
doing their leg kick, showing a simultaneous extension of
their two pairs of limbs. However and contrary to what was
expected, they did not pull with their arms while recovering their legs. This seems more a characteristic of novice
children8 than more mature swimmers.
In a former study, Saito9 measured the stroking characteristics of 163 beginners aged 15–16 y in breaststroke
swimming. After a 6-week period of training, the swimmers performed at a higher speed and this was related to
an improvement of their SL, as their SR did not change. This
evolution was attributed to an improvement of the arm and
leg movements in combination. This coordination aspect will
be discussed hereafter.
5.1.1. Glide vs. motor continuity
The T1 parameter, which characterized the motor and
propulsive continuity after the leg kick, was always positive in
the recreational group. This indicates that the arm propulsive
phase started before the end of the leg in-sweep. By contrast, the competitive group used a gliding coordination at
slow speed. It has also been observed that elite breaststrokers
use glide coordination at sub-maximal speed.4,10 Meanwhile,
they are able to maintain a higher intra-stroke velocity than
less-competitive swimmers during this non-propulsive phase,
just before the arm could take over.11,12
5.1.2. Recreational swimmers have longer arm recovery
The two groups of swimmers spent the same amount
of time recovering their legs, starting at the same moment.
However, the recreational swimmers’ arms took far longer
to recover. This may be linked to a longer breathing time.
Swimming educators routinely observed that beginners can
have an incomplete expiration during the propulsive phase
of the arm, which forces them to finish expiring when their
mouth emerges, just before taking their breath. In total, the
recreational group spent nearly 50% of one stroke cycle in
H. Leblanc et al. / Journal of Science and Medicine in Sport 12 (2009) 352–356
the arm recovery, i.e. in a decelerating phase, as opposed to
29.4% in the competitive group. As a direct consequence, an
overlapping was noticed between the arm recovery and the
leg extension as the negative value of T3 parameter shows.
Further analysis showed that the mean angle formed by the
swimmers’ arms and forearms at the end of the leg recovery
was of 95.1◦ in the recreational swimmers as compared to
156.4◦ for the competitive swimmers (supplementary Figs. 1
and 2 for typical examples).
In recreational swimmers, the end of the extension of the
lower and upper limbs was simultaneous. Some authors13,14
have suggested that the central nervous system imposes two
kinds of constraints in the organization of inter-limb coordination. In the first kind (so-called egocentric), the limbs are
naturally drawn to perform a synchronized motion toward or
away from the longitudinal axis of the body. This seemed
to govern the simultaneous arm and leg extension that was
observed in seven out of twelve of the recreational swimmers showing a symmetrical pattern between the arms and
the legs with respect to the longitudinal axis of their body
(supplementary Fig. 3). In this mirror-like pattern, these
swimmers extended their arms forward and outward, diagonally, along with the leg outward and backward extension.
The six other swimmers extended their arms in a joined position, featuring more a transversal arm–leg symmetry.
In the second kind of constraints (so-called allocentric),
non-homologous limbs (arms vs. legs) tend to have a synchronized motion in the same direction.13,14 That could explain
why all the swimmers synchronized their leg and arm respective recoveries: the flexion of the legs was associated with the
extension of the arms, with both movements directed forward with respect to the swimmers’ body. In the case of
recreational swimmers, however, this constraint seemed to
be overcome by the egocentric constraint in the second half
of the movement.
5.1.3. At a given speed, recreational as competitors used
the same timing to perform their leg and arm propulsive
actions
The direct consequence of the absence of glide phase in
recreational swimmers was that the proportion they devoted
to the propulsive and recovery parts of their arm and leg was
greater in comparison with those of competitors.
The same timing was used by all swimmers to perform
their leg propulsive phase, but competitors could take advantage of the thrust produced by their leg kick because their arms
were already recovered and well aligned. In contrast, recreational swimmers could barely perform any glide because
they had already done their leg kick when this glide should
have appeared. Recreational as competitive swimmers spent
the same amount of time to complete their arm propulsive
phase. But this does not necessarily mean that recreational
swimmers did effectively apply propulsive forces in the water.
It has been already observed that for a given arm-stroke
duration, elite swimmers are able to produce more acceleration of their body and to cover a longer distance during
355
this phase.11,12 This may contribute to the larger SL of elite
swimmers who covered 0.50 m more distance per stroke.
Finally, no differences were found in the relative duration of the propulsive index, but in real terms, this index was
smaller in the recreational swimmers, as a direct consequence
of their overlapped coordination. By doing so, the swimmers
performed two opposite actions altogether: the arm recovery
which occurred underwater and opposes the forward progression of the swimmer; and the leg extension which causes a
velocity increase.1,11,12 Given that, the recreational swimmers could barely perform an efficient leg propulsive action.
This may account for the higher SR they used to maintain the
same speed as competitive swimmers.
5.2. Effect of speed increase on coordination
The second hypothesis was confirmed: having no glide
time, recreational swimmers can hardly modulate their coordination according to the swimming speeds.
5.2.1. Recreational swimmers overlapped their
propulsive phases
At sprint speed, T1 decreased both in percentage and
in real time. From an overlapped coordination, recreational
swimmers moved even more deeply in this type of coordination. An exaggerated overlap time led the swimmers to
start their arm propulsion while their leg kick was still accelerating their body. As a result, they might have increased
their form drag because their arms were no longer streamlined. Form or pressure drag constitutes a resistive force and
forms when water has to be moved away from the swimmer’s body as it progresses forward.7 This overlapped time
of the recreational swimmers is linked to the shortening of
their arm recovery and causes an earlier catch up. The video
replay often shows a partial extension of the arm during the
end of the recovery at sprint pace. The resistance they faced
could even have worsened as they tried to recover their arm as
fast as possible at sprint pace. In swimming, it is well established that a body must overcome a resistive force (drag) that
increases with its squared velocity. Water is some 820 times
more dense than air, and the specificity of this medium may
have a strong impact on the coordination. At higher speeds,
breaststroke swimmers must increase their propulsive force
while minimizing the drag they oppose to motion. This is
quite a difficult task because an increasing propulsive force
will result in an increase of resistive force. This is particularly acute in the breaststroke where all the recovery phases
are performed underwater. This could constrain swimmers of
different skill levels to adopt an overlapped coordination.
5.2.2. Recreational swimmers face difficulties to adapt
their coordination
With the speed increase, competitors switched from a
glide to an overlapped coordination while the other swimmers
seemed “locked” in the overlapped mode. This is consistent
356
H. Leblanc et al. / Journal of Science and Medicine in Sport 12 (2009) 352–356
with the results obtained by Soares et al.3 who did not note
any change in relative arm and leg phase durations of swimmers under different speed conditions. To reinforce this, the
relative value of the propulsive index did not evolve in the
recreational group, whereas it increased in the competitive
group as a result of the decrease of the T1 (glide) parameter.
In this study, beginners somehow were not able to adjust their
coordination according to the pace. Paradoxically, by adopting an overlapped coordination, beginners seem to conform to
a mechanical principle: eliminating the dead space in propulsion in order to limit the decrease in intra-stroke velocity.
Nevertheless, this was not associated with a longer SL or a
higher speed as was the case in the competitive group.
The observation of the coordination pattern of recreational
swimmers reflects a common view of teaching methods
where the arm recovery is represented as occurring during the
leg extension. If the teaching process does start on the basis
of the motor repertoire of the subject, it should not confine
him to a coordination that could impede further progress.
Practical implications
• Swimming educators should include coordination and
glide drills in their programs to improve the arm–leg coordination of recreational swimmers.
• Swimmers should practice coordination drills, (e.g. two
kicks to one arm pull breaststroke), gliding drills (e.g. kick
only breaststroke), and arm pull and recovery drills (e.g.
vertical arm pull or breaststroke pull with flutter kick).
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/j.jsams.
2008.01.001.
References
1. Maglischo EW. Swimming fastest. Champaign, IL: Human Kinetics
Publishing; 2003.
2. Sanders RH. Breaststroke technique variations among New Zealand
Pan Pacific squad swimmers. In: Troup JP, Hollander AP, Strass D,
Trappe SW, Cappaert JM, Trappe TA, editors. Swimming science VII.
London: E & FN SPON; 1996. p. 64–9.
3. Soares PM, Sousa F, Vilas-Boas JP. Differences in breaststroke synchronisation induced by different race velocities. In: Keskinen KL, Komi
PV, Hollander AP, editors. Swimming science VIII. Finland: Jyvaskyla;
1999. p. 53–7.
4. Takagi H, Sugimoto S, Nishijima N, et al. Differences in stroke phases,
arm–leg coordination and velocity fluctuation due to event, gender and
performance level in breaststroke. Sports Biomech 2004;3(1):15–27.
5. Seifert L, Chollet D. A new index of flat breaststroke propulsion:
comparison between elite males and elite women. J Sports Sci
2005;23(3):309–20.
6. Tokuyama H, Okamoto T, Kumamoto M. Electromyographic study of
swimming in infants and children. In: Komi PV, editor. Biomechanics
V-B., international series on biomechanics, vol. IB. Baltimore: Univ
Park Press; 1976. p. 215–21.
7. Wilson B, Thorp R. Active drag in swimming. In: Chatard JC, editor.
Swimming science IX. France: Saint Etienne; 2003. p. 15–21.
8. Taguchi M. Biomechanical analysis of arms and legs coordination in
the breaststroke. Fukuoka Univ Rev Phys Educ 1975;5:181–92.
9. Saito M. The effects of training on the relationships among velocity,
stroke rate and distance per stroke in untrained subjects swimming the
breaststroke. Res Quart Ex Sports 1982;53(4):323–9.
10. Leblanc H, Seifert L, Baudry L, et al. Arm–leg coordination in flat
breaststroke: a comparative study between elite and non-elite swimmers. Int J Sports Med 2005;26(9):787–97.
11. D’Acquisto LJ, Costill DL, Gehlsen GM, et al. Breaststroke economy, skill, and performance: study of breaststroke mechanics using a
computer based “velocity-video” system. J Swim Res 1988;4(2):9–13.
12. Leblanc H, Seifert L, Tourny-Chollet C, et al. Intra-cyclic distance
per stroke phase, velocity fluctuations and acceleration time ratio of a
breaststroker’s hip: a comparison between elite and non-elite swimmers
at different race paces. Int J Sports Med 2007;28(2):140–7.
13. Swinnen SP, Jardin K, Meulenbroek R, et al. Egocentric and allocentric
constraints in the expression of patterns of inter-limb coordination. J
Cogn Neurosci 1997;9:348–77.
14. Baldissera F, Cavalleri P, Civaschi P. Preferential coupling between voluntary movements of ipsilateral limbs. Neurosci Lett 1982;74:95–100.